sakman has quit [Remote host closed the connection]
frieder_ has joined #u-boot
frieder_ has quit [Remote host closed the connection]
sakman has joined #u-boot
monstr_ has joined #u-boot
monstr has quit [Ping timeout: 256 seconds]
frieder has quit [Ping timeout: 246 seconds]
monstr_ has quit [Ping timeout: 250 seconds]
frieder has joined #u-boot
naoki has quit [Quit: naoki]
valdemaras has joined #u-boot
sng has quit [Remote host closed the connection]
sukrutb has quit [Ping timeout: 245 seconds]
flyback has joined #u-boot
prabhakarlad has quit [Quit: Client closed]
<paulbarker>
Hi folks. Is there any documentation for how `mem_map` is used on arm64 devices (declared in arch/arm/include/asm/armv8/mmu.h)? Or what the `PTE_*` and `MT_*` symbols mean?
prabhakarlad has joined #u-boot
mmu_man has quit [Ping timeout: 246 seconds]
<Sout_>
while not that helpful, i assume MT == memory type from the documentation, and PTE == page table entry?
<rfs613>
paulbarker: yes the ARM docs are certainly good reference... and they take a long time to digest...
<paulbarker>
It's starting to make a bit of sense
<paulbarker>
And looking at the code, `mem_map` is just used to set up the TLB
elpogo has joined #u-boot
<marex>
paulbarker: I also recall a talk at ELCE (?) from uh ... one of the ARM people ... about formal memory model for kernel
<paulbarker>
marex: I'll do a bit of a google
<elpogo>
hello. is there any way to get the u-boot prompt to show up on qemu's virtconsole device? i'm using u-boot.bin from guix's u-boot-qemu-arm64 package
<paulbarker>
marex: The context here is I need to update `arch/arm/mach-rmobile/memmap-gen3.c` to support the RZ/G2L family, which has a slightly different memory map.
Leopold has joined #u-boot
<paulbarker>
And I don't want to just copy paste things without understanding what the defines mean
<marex>
ah
rfs613 has quit [Ping timeout: 246 seconds]
rfs613 has joined #u-boot
<paulbarker>
marex: I think I now just need to understand what PTE_BLOCK_INNER_SHARE means
<sng>
and I have been able to make changes to move the capsule generation through binman to the test/py/tests/test_efi_capsule/capsule_gen_binman.dts file for the tests. so i was wondering if I should be sending a V11 series which makes this change?
<sng>
since the series has not been merged yet
<marex>
Clamor: imx also uses chipidea ehci + PHY
<sng>
Tom had earlier suggested sending a follow-up series. but since I have this working, and the series has not been merged, I was wondering if I should send a next version of the series, so that this gets addressed before getting merged
<Clamor>
marex: may you send me a link to this bundle
valdemaras has joined #u-boot
<NishanthMenon>
sjg1: https://github.com/u-boot/u-boot/blob/master/include/env_default.h#L115 -> Any reason for env file to be overwriten by comon CFG_EXTRA_ENV_SETTINGS ? I am trying to set the board/xyz.env to change boot_targets=mmc0 mmc1 usb pxe set in my common config.h file to boot_targets=mmc1 mmc0 usb pxe and it does'nt work..wondering if there is a good reason I am missing
<sng>
so that all the patches can be applied together. the series for embedding the public key for capsule authentication now has a r-b on all relevant patches from Tartarus
<sng>
Tartarus: hope that is fine with you
<Tartarus>
sounds good, thanks
mmu_man has joined #u-boot
prabhakarlad has quit [Quit: Client closed]
LeSpocky has quit [Quit: reboot]
<sjg1>
sng: Tartarus: I think it is best to let your first series be merged. Perhaps Tom might want to take them both at the same time, not sure. But the series are logically separate IMO
<sjg1>
sng: Tartarus Oh just saw the reply above...well...I don't want to review another version of the first series, but it sounds like Tom will
<sjg1>
Tartarus: What happened to the ipv6 series?
<marex>
Clamor: why ?
<sng>
sjg1: Tartarus: Simon, there is only one patch which undergoes a change. that is moving the binman nodes for capsule generation under the test directory. the rest of the patches are the same
<kwizart>
Clamor, your tree boots on ac100 but there fastboot isn't enabled (to test fastboot usb as requested) and flash_uboot failed with error finding u-boot-dtb-tegra.bin which isn't surprising as ls mmc dev 0:1 outputs my /boot partition (/ is f2fs formatted, so can't list)
<sng>
Tartarus: sjg1: regarding clubbing of the two series (capsule gen + pub key embed), I am fine with whatever approach that suits you and Tom
<Clamor>
marex: stuff you have sent is irrelevant apart imx one I suppose. Why? Because tegra ehci is a wild mix of controller and phy involving clock setups, reset setups and a ton of registers configuration for all tegra generations. One wrong move and all this glass store crashes.
<marex>
Clamor: maybe tagr can help test it all ?
<Clamor>
marex: I have said, I already have too much stuff to upstream, I would like to handle those first. And then I might return to ehci
<Clamor>
kwizart: as always, too much knowledge makes testers bad ;)
<Clamor>
But non the less thanks, good to know that it boots
foxtrot has quit [Remote host closed the connection]
foxtrot has joined #u-boot
<sjg1>
sng: Me too, but I won't be reviewing the first one again. There were a lot of revisions and I suspect we are both exhasusted
ikarso has joined #u-boot
wkennington has joined #u-boot
<marex>
Clamor: take your time, I know the feeling /wrt walls of patches
<wkennington>
I'm trying to debug an issue with an rk3588 board where when I load a fitImage of larger than 4MB it fails to validate it when trying to `bootm`
<wkennington>
any pointers on what config values to check to reason about why it fails to validate?
<wkennington>
i figure it's 2 overlapping memory regions
<Clamor>
marex: thanks
<marex>
wkennington: look at the load address embedded in the fitImage image/ node entries , maybe you are overwriting something ?
<marex>
wkennington: compare it with fitImage loadaddress (the one you are passing to bootm)
rvalue has quit [Read error: Connection reset by peer]
vagrantc has joined #u-boot
<wkennington>
marex: wouldn't it validate the fitimage before loading?
<wkennington>
i feel like the fitimage printout would start printing the first entry
rvalue has joined #u-boot
sng_ has joined #u-boot
sng has quit [Read error: Connection reset by peer]
<wkennington>
yeah, this is just an image for bringup
<wkennington>
i need to plumb some metadata for stored images in the flash
<marex>
why not use a filesystem ?
<wkennington>
i've been wanting to avoid using a vfat to do a static a/b setup
<marex>
use ext4 ?
<wkennington>
i mean, i want guarantees neither file can exceed certain boundsa
<wkennington>
so one bad update couldn't prevent another
<marex>
wkennington: like ... filesystem indicates it is full kind of thing ?
<wkennington>
yeah, although i could have checks in the OS to prevent writing out files that are too big
<marex>
wkennington: if you are doing A/B updates, then just have two /boot and two /root partitions
<wkennington>
sure, but a /boot with a single fitimage file is kinda dumb
<marex>
wkennington: you dont have to write the full partition image, although it is probably a good idea with boot partition to make sure the filesystem is always consistent
<marex>
wkennington: why ?
<marex>
wkennington: it gives you the "metadata" part
<wkennington>
yeah, it's just silly
<marex>
wkennington: you can also stick a fitImage into a partition, then read the whole partition in U-Boot and you would be sure you did read enough
<wkennington>
that was my plan
<marex>
wkennington: or you can read the fitImage header, determine the size, and then read the whole thing
<wkennington>
yep
<wkennington>
this was my plan
<wkennington>
but i just wanted to get the kernel booting before i hacked that in
<marex>
I'd probably just opt for as standard and as wildly used (=tested) option available
prabhakarlad has joined #u-boot
<wkennington>
is there any way to store the stdout result of some commands to a var?
<wkennington>
i guess most command output is just for debugging
<marex>
wkennington: nope
<marex>
at least not yet
<marex>
wkennington: why ?
<wkennington>
just wondering for scripting sake
<marex>
wkennington: what are you trying to script ?